gokulsrinivasagan
commited on
Commit
•
d621c37
1
Parent(s):
67714db
Model save
Browse files- README.md +413 -0
- config.json +39 -0
- generation_config.json +6 -0
- logs/events.out.tfevents.1719842149.s_005_m.77188.0 +3 -0
- merges.txt +0 -0
- model.safetensors +3 -0
- special_tokens_map.json +5 -0
- tokenizer.json +0 -0
- tokenizer_config.json +19 -0
- training_args.bin +3 -0
- vocab.json +0 -0
README.md
ADDED
@@ -0,0 +1,413 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: mit
|
3 |
+
base_model: openai-community/gpt2
|
4 |
+
tags:
|
5 |
+
- generated_from_trainer
|
6 |
+
metrics:
|
7 |
+
- accuracy
|
8 |
+
model-index:
|
9 |
+
- name: gpt_train_12_128
|
10 |
+
results: []
|
11 |
+
---
|
12 |
+
|
13 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
14 |
+
should probably proofread and complete it, then remove this comment. -->
|
15 |
+
|
16 |
+
# gpt_train_12_128
|
17 |
+
|
18 |
+
This model is a fine-tuned version of [openai-community/gpt2](https://huggingface.co/openai-community/gpt2) on an unknown dataset.
|
19 |
+
It achieves the following results on the evaluation set:
|
20 |
+
- Loss: 10.0781
|
21 |
+
- Accuracy: 0.0781
|
22 |
+
|
23 |
+
## Model description
|
24 |
+
|
25 |
+
More information needed
|
26 |
+
|
27 |
+
## Intended uses & limitations
|
28 |
+
|
29 |
+
More information needed
|
30 |
+
|
31 |
+
## Training and evaluation data
|
32 |
+
|
33 |
+
More information needed
|
34 |
+
|
35 |
+
## Training procedure
|
36 |
+
|
37 |
+
### Training hyperparameters
|
38 |
+
|
39 |
+
The following hyperparameters were used during training:
|
40 |
+
- learning_rate: 1e-05
|
41 |
+
- train_batch_size: 48
|
42 |
+
- eval_batch_size: 48
|
43 |
+
- seed: 10
|
44 |
+
- distributed_type: multi-GPU
|
45 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
46 |
+
- lr_scheduler_type: linear
|
47 |
+
- num_epochs: 100
|
48 |
+
- mixed_precision_training: Native AMP
|
49 |
+
|
50 |
+
### Training results
|
51 |
+
|
52 |
+
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
53 |
+
|:-------------:|:------:|:----:|:---------------:|:--------:|
|
54 |
+
| 10.8438 | 0.0001 | 1 | 10.8438 | 0.0103 |
|
55 |
+
| 10.8359 | 0.0001 | 2 | 10.8438 | 0.0103 |
|
56 |
+
| 10.8438 | 0.0002 | 3 | 10.8438 | 0.0103 |
|
57 |
+
| 10.8359 | 0.0003 | 4 | 10.8438 | 0.0103 |
|
58 |
+
| 10.8438 | 0.0004 | 5 | 10.8438 | 0.0103 |
|
59 |
+
| 10.8438 | 0.0004 | 6 | 10.8438 | 0.0103 |
|
60 |
+
| 10.8359 | 0.0005 | 7 | 10.8438 | 0.0103 |
|
61 |
+
| 10.8359 | 0.0006 | 8 | 10.8438 | 0.0103 |
|
62 |
+
| 10.8438 | 0.0007 | 9 | 10.8438 | 0.0103 |
|
63 |
+
| 10.8359 | 0.0007 | 10 | 10.8438 | 0.0103 |
|
64 |
+
| 10.8359 | 0.0008 | 11 | 10.8438 | 0.0103 |
|
65 |
+
| 10.8438 | 0.0009 | 12 | 10.8438 | 0.0103 |
|
66 |
+
| 10.8438 | 0.0009 | 13 | 10.8438 | 0.0103 |
|
67 |
+
| 10.8359 | 0.0010 | 14 | 10.8438 | 0.0103 |
|
68 |
+
| 10.8359 | 0.0011 | 15 | 10.8438 | 0.0103 |
|
69 |
+
| 10.8438 | 0.0012 | 16 | 10.8438 | 0.0103 |
|
70 |
+
| 10.8359 | 0.0012 | 17 | 10.8438 | 0.0103 |
|
71 |
+
| 10.8438 | 0.0013 | 18 | 10.8281 | 0.0113 |
|
72 |
+
| 10.8203 | 0.0014 | 19 | 10.8125 | 0.0116 |
|
73 |
+
| 10.8203 | 0.0015 | 20 | 10.8047 | 0.0117 |
|
74 |
+
| 10.8047 | 0.0015 | 21 | 10.7891 | 0.0118 |
|
75 |
+
| 10.7969 | 0.0016 | 22 | 10.7734 | 0.0118 |
|
76 |
+
| 10.7812 | 0.0017 | 23 | 10.7656 | 0.0118 |
|
77 |
+
| 10.7656 | 0.0017 | 24 | 10.75 | 0.0119 |
|
78 |
+
| 10.7578 | 0.0018 | 25 | 10.7344 | 0.0121 |
|
79 |
+
| 10.75 | 0.0019 | 26 | 10.7266 | 0.0124 |
|
80 |
+
| 10.7344 | 0.0020 | 27 | 10.7188 | 0.0131 |
|
81 |
+
| 10.7266 | 0.0020 | 28 | 10.7031 | 0.0144 |
|
82 |
+
| 10.7109 | 0.0021 | 29 | 10.6953 | 0.0165 |
|
83 |
+
| 10.7031 | 0.0022 | 30 | 10.6875 | 0.0196 |
|
84 |
+
| 10.7031 | 0.0023 | 31 | 10.6797 | 0.0236 |
|
85 |
+
| 10.6875 | 0.0023 | 32 | 10.6719 | 0.0282 |
|
86 |
+
| 10.6797 | 0.0024 | 33 | 10.6641 | 0.0330 |
|
87 |
+
| 10.6719 | 0.0025 | 34 | 10.6641 | 0.0375 |
|
88 |
+
| 10.6719 | 0.0025 | 35 | 10.6562 | 0.0409 |
|
89 |
+
| 10.6719 | 0.0026 | 36 | 10.6484 | 0.0437 |
|
90 |
+
| 10.6484 | 0.0027 | 37 | 10.6484 | 0.0461 |
|
91 |
+
| 10.6562 | 0.0028 | 38 | 10.6406 | 0.0481 |
|
92 |
+
| 10.6484 | 0.0028 | 39 | 10.6406 | 0.0498 |
|
93 |
+
| 10.6484 | 0.0029 | 40 | 10.6328 | 0.0511 |
|
94 |
+
| 10.6406 | 0.0030 | 41 | 10.6328 | 0.0521 |
|
95 |
+
| 10.6406 | 0.0031 | 42 | 10.625 | 0.0529 |
|
96 |
+
| 10.6406 | 0.0031 | 43 | 10.625 | 0.0535 |
|
97 |
+
| 10.625 | 0.0032 | 44 | 10.6172 | 0.0539 |
|
98 |
+
| 10.625 | 0.0033 | 45 | 10.6172 | 0.0542 |
|
99 |
+
| 10.625 | 0.0033 | 46 | 10.6172 | 0.0543 |
|
100 |
+
| 10.6172 | 0.0034 | 47 | 10.6094 | 0.0544 |
|
101 |
+
| 10.625 | 0.0035 | 48 | 10.6094 | 0.0545 |
|
102 |
+
| 10.6172 | 0.0036 | 49 | 10.6016 | 0.0545 |
|
103 |
+
| 10.6016 | 0.0036 | 50 | 10.6016 | 0.0545 |
|
104 |
+
| 10.6016 | 0.0037 | 51 | 10.6016 | 0.0545 |
|
105 |
+
| 10.6016 | 0.0038 | 52 | 10.5938 | 0.0546 |
|
106 |
+
| 10.6016 | 0.0039 | 53 | 10.5938 | 0.0545 |
|
107 |
+
| 10.5938 | 0.0039 | 54 | 10.5938 | 0.0545 |
|
108 |
+
| 10.6016 | 0.0040 | 55 | 10.5859 | 0.0545 |
|
109 |
+
| 10.5859 | 0.0041 | 56 | 10.5859 | 0.0545 |
|
110 |
+
| 10.6016 | 0.0041 | 57 | 10.5859 | 0.0545 |
|
111 |
+
| 10.5859 | 0.0042 | 58 | 10.5859 | 0.0546 |
|
112 |
+
| 10.5859 | 0.0043 | 59 | 10.5781 | 0.0547 |
|
113 |
+
| 10.5781 | 0.0044 | 60 | 10.5781 | 0.0548 |
|
114 |
+
| 10.5781 | 0.0044 | 61 | 10.5781 | 0.0550 |
|
115 |
+
| 10.5781 | 0.0045 | 62 | 10.5703 | 0.0553 |
|
116 |
+
| 10.5781 | 0.0046 | 63 | 10.5703 | 0.0557 |
|
117 |
+
| 10.5703 | 0.0046 | 64 | 10.5703 | 0.0561 |
|
118 |
+
| 10.5781 | 0.0047 | 65 | 10.5625 | 0.0566 |
|
119 |
+
| 10.5625 | 0.0048 | 66 | 10.5625 | 0.0570 |
|
120 |
+
| 10.5781 | 0.0049 | 67 | 10.5625 | 0.0573 |
|
121 |
+
| 10.5703 | 0.0049 | 68 | 10.5547 | 0.0575 |
|
122 |
+
| 10.5625 | 0.0050 | 69 | 10.5547 | 0.0577 |
|
123 |
+
| 10.5625 | 0.0051 | 70 | 10.5547 | 0.0578 |
|
124 |
+
| 10.5625 | 0.0052 | 71 | 10.5547 | 0.0579 |
|
125 |
+
| 10.5547 | 0.0052 | 72 | 10.5469 | 0.0580 |
|
126 |
+
| 10.5469 | 0.0053 | 73 | 10.5469 | 0.0580 |
|
127 |
+
| 10.5469 | 0.0054 | 74 | 10.5469 | 0.0580 |
|
128 |
+
| 10.5547 | 0.0054 | 75 | 10.5391 | 0.0580 |
|
129 |
+
| 10.5547 | 0.0055 | 76 | 10.5391 | 0.0580 |
|
130 |
+
| 10.5469 | 0.0056 | 77 | 10.5391 | 0.0582 |
|
131 |
+
| 10.5469 | 0.0057 | 78 | 10.5391 | 0.0582 |
|
132 |
+
| 10.5312 | 0.0057 | 79 | 10.5312 | 0.0584 |
|
133 |
+
| 10.5312 | 0.0058 | 80 | 10.5312 | 0.0586 |
|
134 |
+
| 10.5312 | 0.0059 | 81 | 10.5312 | 0.0590 |
|
135 |
+
| 10.5312 | 0.0060 | 82 | 10.5312 | 0.0593 |
|
136 |
+
| 10.5312 | 0.0060 | 83 | 10.5234 | 0.0597 |
|
137 |
+
| 10.5234 | 0.0061 | 84 | 10.5234 | 0.0600 |
|
138 |
+
| 10.5312 | 0.0062 | 85 | 10.5234 | 0.0602 |
|
139 |
+
| 10.5312 | 0.0062 | 86 | 10.5234 | 0.0603 |
|
140 |
+
| 10.5234 | 0.0063 | 87 | 10.5156 | 0.0604 |
|
141 |
+
| 10.5156 | 0.0064 | 88 | 10.5156 | 0.0605 |
|
142 |
+
| 10.5234 | 0.0065 | 89 | 10.5156 | 0.0606 |
|
143 |
+
| 10.5156 | 0.0065 | 90 | 10.5156 | 0.0606 |
|
144 |
+
| 10.5156 | 0.0066 | 91 | 10.5078 | 0.0606 |
|
145 |
+
| 10.5156 | 0.0067 | 92 | 10.5078 | 0.0605 |
|
146 |
+
| 10.5156 | 0.0068 | 93 | 10.5078 | 0.0603 |
|
147 |
+
| 10.5156 | 0.0068 | 94 | 10.5078 | 0.0602 |
|
148 |
+
| 10.5234 | 0.0069 | 95 | 10.5 | 0.0601 |
|
149 |
+
| 10.5156 | 0.0070 | 96 | 10.5 | 0.0602 |
|
150 |
+
| 10.5078 | 0.0070 | 97 | 10.5 | 0.0603 |
|
151 |
+
| 10.5 | 0.0071 | 98 | 10.5 | 0.0603 |
|
152 |
+
| 10.5078 | 0.0072 | 99 | 10.5 | 0.0604 |
|
153 |
+
| 10.5078 | 0.0073 | 100 | 10.4922 | 0.0606 |
|
154 |
+
| 10.5 | 0.0073 | 101 | 10.4922 | 0.0607 |
|
155 |
+
| 10.4922 | 0.0074 | 102 | 10.4922 | 0.0609 |
|
156 |
+
| 10.4922 | 0.0075 | 103 | 10.4922 | 0.0612 |
|
157 |
+
| 10.4844 | 0.0076 | 104 | 10.4844 | 0.0614 |
|
158 |
+
| 10.4922 | 0.0076 | 105 | 10.4844 | 0.0617 |
|
159 |
+
| 10.4922 | 0.0077 | 106 | 10.4844 | 0.0619 |
|
160 |
+
| 10.4844 | 0.0078 | 107 | 10.4844 | 0.0622 |
|
161 |
+
| 10.4922 | 0.0078 | 108 | 10.4766 | 0.0625 |
|
162 |
+
| 10.4844 | 0.0079 | 109 | 10.4766 | 0.0628 |
|
163 |
+
| 10.4766 | 0.0080 | 110 | 10.4766 | 0.0630 |
|
164 |
+
| 10.4844 | 0.0081 | 111 | 10.4766 | 0.0632 |
|
165 |
+
| 10.4766 | 0.0081 | 112 | 10.4766 | 0.0634 |
|
166 |
+
| 10.4844 | 0.0082 | 113 | 10.4688 | 0.0636 |
|
167 |
+
| 10.4766 | 0.0083 | 114 | 10.4688 | 0.0638 |
|
168 |
+
| 10.4766 | 0.0084 | 115 | 10.4688 | 0.0640 |
|
169 |
+
| 10.4844 | 0.0084 | 116 | 10.4688 | 0.0643 |
|
170 |
+
| 10.4531 | 0.0085 | 117 | 10.4609 | 0.0644 |
|
171 |
+
| 10.4609 | 0.0086 | 118 | 10.4609 | 0.0647 |
|
172 |
+
| 10.4609 | 0.0086 | 119 | 10.4609 | 0.0648 |
|
173 |
+
| 10.4688 | 0.0087 | 120 | 10.4609 | 0.0649 |
|
174 |
+
| 10.4609 | 0.0088 | 121 | 10.4609 | 0.0651 |
|
175 |
+
| 10.4609 | 0.0089 | 122 | 10.4531 | 0.0653 |
|
176 |
+
| 10.4531 | 0.0089 | 123 | 10.4531 | 0.0656 |
|
177 |
+
| 10.4531 | 0.0090 | 124 | 10.4531 | 0.0659 |
|
178 |
+
| 10.4531 | 0.0091 | 125 | 10.4531 | 0.0660 |
|
179 |
+
| 10.4531 | 0.0092 | 126 | 10.4453 | 0.0662 |
|
180 |
+
| 10.4531 | 0.0092 | 127 | 10.4453 | 0.0664 |
|
181 |
+
| 10.4453 | 0.0093 | 128 | 10.4453 | 0.0667 |
|
182 |
+
| 10.4531 | 0.0094 | 129 | 10.4453 | 0.0670 |
|
183 |
+
| 10.4375 | 0.0094 | 130 | 10.4453 | 0.0673 |
|
184 |
+
| 10.4453 | 0.0095 | 131 | 10.4375 | 0.0676 |
|
185 |
+
| 10.4375 | 0.0096 | 132 | 10.4375 | 0.0678 |
|
186 |
+
| 10.4375 | 0.0097 | 133 | 10.4375 | 0.0679 |
|
187 |
+
| 10.4297 | 0.0097 | 134 | 10.4375 | 0.0679 |
|
188 |
+
| 10.4453 | 0.0098 | 135 | 10.4297 | 0.0678 |
|
189 |
+
| 10.4375 | 0.0099 | 136 | 10.4297 | 0.0677 |
|
190 |
+
| 10.4375 | 0.0100 | 137 | 10.4297 | 0.0677 |
|
191 |
+
| 10.4219 | 0.0100 | 138 | 10.4297 | 0.0677 |
|
192 |
+
| 10.4375 | 0.0101 | 139 | 10.4219 | 0.0678 |
|
193 |
+
| 10.4297 | 0.0102 | 140 | 10.4219 | 0.0680 |
|
194 |
+
| 10.4297 | 0.0102 | 141 | 10.4219 | 0.0682 |
|
195 |
+
| 10.4219 | 0.0103 | 142 | 10.4219 | 0.0684 |
|
196 |
+
| 10.4219 | 0.0104 | 143 | 10.4219 | 0.0687 |
|
197 |
+
| 10.4219 | 0.0105 | 144 | 10.4141 | 0.0689 |
|
198 |
+
| 10.4219 | 0.0105 | 145 | 10.4141 | 0.0692 |
|
199 |
+
| 10.4141 | 0.0106 | 146 | 10.4141 | 0.0693 |
|
200 |
+
| 10.4062 | 0.0107 | 147 | 10.4141 | 0.0695 |
|
201 |
+
| 10.4141 | 0.0108 | 148 | 10.4062 | 0.0696 |
|
202 |
+
| 10.4141 | 0.0108 | 149 | 10.4062 | 0.0697 |
|
203 |
+
| 10.4219 | 0.0109 | 150 | 10.4062 | 0.0697 |
|
204 |
+
| 10.4062 | 0.0110 | 151 | 10.4062 | 0.0698 |
|
205 |
+
| 10.4141 | 0.0110 | 152 | 10.4062 | 0.0700 |
|
206 |
+
| 10.4141 | 0.0111 | 153 | 10.3984 | 0.0701 |
|
207 |
+
| 10.4219 | 0.0112 | 154 | 10.3984 | 0.0702 |
|
208 |
+
| 10.4141 | 0.0113 | 155 | 10.3984 | 0.0704 |
|
209 |
+
| 10.4062 | 0.0113 | 156 | 10.3984 | 0.0705 |
|
210 |
+
| 10.4062 | 0.0114 | 157 | 10.3906 | 0.0707 |
|
211 |
+
| 10.3906 | 0.0115 | 158 | 10.3906 | 0.0708 |
|
212 |
+
| 10.3906 | 0.0116 | 159 | 10.3906 | 0.0710 |
|
213 |
+
| 10.3984 | 0.0116 | 160 | 10.3906 | 0.0711 |
|
214 |
+
| 10.3984 | 0.0117 | 161 | 10.3906 | 0.0711 |
|
215 |
+
| 10.3906 | 0.0118 | 162 | 10.3828 | 0.0712 |
|
216 |
+
| 10.3906 | 0.0118 | 163 | 10.3828 | 0.0712 |
|
217 |
+
| 10.3906 | 0.0119 | 164 | 10.3828 | 0.0714 |
|
218 |
+
| 10.3828 | 0.0120 | 165 | 10.3828 | 0.0715 |
|
219 |
+
| 10.375 | 0.0121 | 166 | 10.375 | 0.0716 |
|
220 |
+
| 10.3828 | 0.0121 | 167 | 10.375 | 0.0717 |
|
221 |
+
| 10.3828 | 0.0122 | 168 | 10.375 | 0.0718 |
|
222 |
+
| 10.3828 | 0.0123 | 169 | 10.375 | 0.0719 |
|
223 |
+
| 10.3828 | 0.0124 | 170 | 10.375 | 0.0721 |
|
224 |
+
| 10.3672 | 0.0124 | 171 | 10.3672 | 0.0721 |
|
225 |
+
| 10.375 | 0.0125 | 172 | 10.3672 | 0.0721 |
|
226 |
+
| 10.3594 | 0.0126 | 173 | 10.3672 | 0.0721 |
|
227 |
+
| 10.375 | 0.0126 | 174 | 10.3672 | 0.0720 |
|
228 |
+
| 10.3594 | 0.0127 | 175 | 10.3594 | 0.0721 |
|
229 |
+
| 10.3672 | 0.0128 | 176 | 10.3594 | 0.0722 |
|
230 |
+
| 10.375 | 0.0129 | 177 | 10.3594 | 0.0723 |
|
231 |
+
| 10.3672 | 0.0129 | 178 | 10.3594 | 0.0726 |
|
232 |
+
| 10.3672 | 0.0130 | 179 | 10.3594 | 0.0727 |
|
233 |
+
| 10.3594 | 0.0131 | 180 | 10.3516 | 0.0728 |
|
234 |
+
| 10.3672 | 0.0132 | 181 | 10.3516 | 0.0729 |
|
235 |
+
| 10.3594 | 0.0132 | 182 | 10.3516 | 0.0730 |
|
236 |
+
| 10.3516 | 0.0133 | 183 | 10.3516 | 0.0731 |
|
237 |
+
| 10.3594 | 0.0134 | 184 | 10.3516 | 0.0732 |
|
238 |
+
| 10.3516 | 0.0134 | 185 | 10.3438 | 0.0733 |
|
239 |
+
| 10.3516 | 0.0135 | 186 | 10.3438 | 0.0733 |
|
240 |
+
| 10.3438 | 0.0136 | 187 | 10.3438 | 0.0734 |
|
241 |
+
| 10.3516 | 0.0137 | 188 | 10.3438 | 0.0734 |
|
242 |
+
| 10.3516 | 0.0137 | 189 | 10.3359 | 0.0735 |
|
243 |
+
| 10.3438 | 0.0138 | 190 | 10.3359 | 0.0735 |
|
244 |
+
| 10.3516 | 0.0139 | 191 | 10.3359 | 0.0735 |
|
245 |
+
| 10.3359 | 0.0139 | 192 | 10.3359 | 0.0737 |
|
246 |
+
| 10.3359 | 0.0140 | 193 | 10.3359 | 0.0737 |
|
247 |
+
| 10.3359 | 0.0141 | 194 | 10.3281 | 0.0736 |
|
248 |
+
| 10.3359 | 0.0142 | 195 | 10.3281 | 0.0736 |
|
249 |
+
| 10.3359 | 0.0142 | 196 | 10.3281 | 0.0736 |
|
250 |
+
| 10.3281 | 0.0143 | 197 | 10.3281 | 0.0737 |
|
251 |
+
| 10.3359 | 0.0144 | 198 | 10.3281 | 0.0738 |
|
252 |
+
| 10.3203 | 0.0145 | 199 | 10.3203 | 0.0740 |
|
253 |
+
| 10.3359 | 0.0145 | 200 | 10.3203 | 0.0741 |
|
254 |
+
| 10.3359 | 0.0146 | 201 | 10.3203 | 0.0742 |
|
255 |
+
| 10.3281 | 0.0147 | 202 | 10.3203 | 0.0743 |
|
256 |
+
| 10.3203 | 0.0147 | 203 | 10.3125 | 0.0743 |
|
257 |
+
| 10.3203 | 0.0148 | 204 | 10.3125 | 0.0743 |
|
258 |
+
| 10.3281 | 0.0149 | 205 | 10.3125 | 0.0743 |
|
259 |
+
| 10.3125 | 0.0150 | 206 | 10.3125 | 0.0741 |
|
260 |
+
| 10.3125 | 0.0150 | 207 | 10.3125 | 0.0740 |
|
261 |
+
| 10.3047 | 0.0151 | 208 | 10.3047 | 0.0740 |
|
262 |
+
| 10.3125 | 0.0152 | 209 | 10.3047 | 0.0741 |
|
263 |
+
| 10.3125 | 0.0153 | 210 | 10.3047 | 0.0742 |
|
264 |
+
| 10.3203 | 0.0153 | 211 | 10.3047 | 0.0743 |
|
265 |
+
| 10.3047 | 0.0154 | 212 | 10.3047 | 0.0744 |
|
266 |
+
| 10.3203 | 0.0155 | 213 | 10.2969 | 0.0745 |
|
267 |
+
| 10.3125 | 0.0155 | 214 | 10.2969 | 0.0747 |
|
268 |
+
| 10.3047 | 0.0156 | 215 | 10.2969 | 0.0749 |
|
269 |
+
| 10.2969 | 0.0157 | 216 | 10.2969 | 0.0750 |
|
270 |
+
| 10.3047 | 0.0158 | 217 | 10.2969 | 0.0750 |
|
271 |
+
| 10.2969 | 0.0158 | 218 | 10.2891 | 0.0749 |
|
272 |
+
| 10.2891 | 0.0159 | 219 | 10.2891 | 0.0747 |
|
273 |
+
| 10.2969 | 0.0160 | 220 | 10.2891 | 0.0744 |
|
274 |
+
| 10.2969 | 0.0161 | 221 | 10.2891 | 0.0742 |
|
275 |
+
| 10.2891 | 0.0161 | 222 | 10.2891 | 0.0741 |
|
276 |
+
| 10.2891 | 0.0162 | 223 | 10.2812 | 0.0742 |
|
277 |
+
| 10.2891 | 0.0163 | 224 | 10.2812 | 0.0743 |
|
278 |
+
| 10.2891 | 0.0163 | 225 | 10.2812 | 0.0746 |
|
279 |
+
| 10.2969 | 0.0164 | 226 | 10.2812 | 0.0748 |
|
280 |
+
| 10.2812 | 0.0165 | 227 | 10.2734 | 0.0749 |
|
281 |
+
| 10.2891 | 0.0166 | 228 | 10.2734 | 0.0750 |
|
282 |
+
| 10.2734 | 0.0166 | 229 | 10.2734 | 0.0751 |
|
283 |
+
| 10.2969 | 0.0167 | 230 | 10.2734 | 0.0750 |
|
284 |
+
| 10.2656 | 0.0168 | 231 | 10.2734 | 0.0749 |
|
285 |
+
| 10.2734 | 0.0169 | 232 | 10.2656 | 0.0747 |
|
286 |
+
| 10.2734 | 0.0169 | 233 | 10.2656 | 0.0747 |
|
287 |
+
| 10.2734 | 0.0170 | 234 | 10.2656 | 0.0746 |
|
288 |
+
| 10.2656 | 0.0171 | 235 | 10.2656 | 0.0747 |
|
289 |
+
| 10.2656 | 0.0171 | 236 | 10.2656 | 0.0748 |
|
290 |
+
| 10.2734 | 0.0172 | 237 | 10.2578 | 0.0749 |
|
291 |
+
| 10.2656 | 0.0173 | 238 | 10.2578 | 0.0752 |
|
292 |
+
| 10.2734 | 0.0174 | 239 | 10.2578 | 0.0755 |
|
293 |
+
| 10.2578 | 0.0174 | 240 | 10.2578 | 0.0756 |
|
294 |
+
| 10.2734 | 0.0175 | 241 | 10.2578 | 0.0756 |
|
295 |
+
| 10.2656 | 0.0176 | 242 | 10.25 | 0.0756 |
|
296 |
+
| 10.2578 | 0.0177 | 243 | 10.25 | 0.0756 |
|
297 |
+
| 10.2578 | 0.0177 | 244 | 10.25 | 0.0756 |
|
298 |
+
| 10.2578 | 0.0178 | 245 | 10.25 | 0.0756 |
|
299 |
+
| 10.2578 | 0.0179 | 246 | 10.25 | 0.0756 |
|
300 |
+
| 10.2578 | 0.0179 | 247 | 10.2422 | 0.0757 |
|
301 |
+
| 10.2578 | 0.0180 | 248 | 10.2422 | 0.0758 |
|
302 |
+
| 10.2422 | 0.0181 | 249 | 10.2422 | 0.0759 |
|
303 |
+
| 10.2422 | 0.0182 | 250 | 10.2422 | 0.0759 |
|
304 |
+
| 10.2422 | 0.0182 | 251 | 10.2422 | 0.0759 |
|
305 |
+
| 10.2422 | 0.0183 | 252 | 10.2344 | 0.0759 |
|
306 |
+
| 10.2422 | 0.0184 | 253 | 10.2344 | 0.0759 |
|
307 |
+
| 10.2422 | 0.0185 | 254 | 10.2344 | 0.0759 |
|
308 |
+
| 10.2422 | 0.0185 | 255 | 10.2344 | 0.0761 |
|
309 |
+
| 10.2422 | 0.0186 | 256 | 10.2344 | 0.0761 |
|
310 |
+
| 10.2422 | 0.0187 | 257 | 10.2266 | 0.0760 |
|
311 |
+
| 10.2422 | 0.0187 | 258 | 10.2266 | 0.0760 |
|
312 |
+
| 10.2344 | 0.0188 | 259 | 10.2266 | 0.0759 |
|
313 |
+
| 10.2344 | 0.0189 | 260 | 10.2266 | 0.0759 |
|
314 |
+
| 10.2266 | 0.0190 | 261 | 10.2266 | 0.0760 |
|
315 |
+
| 10.2188 | 0.0190 | 262 | 10.2188 | 0.0760 |
|
316 |
+
| 10.2266 | 0.0191 | 263 | 10.2188 | 0.0762 |
|
317 |
+
| 10.2266 | 0.0192 | 264 | 10.2188 | 0.0762 |
|
318 |
+
| 10.2188 | 0.0193 | 265 | 10.2188 | 0.0762 |
|
319 |
+
| 10.2266 | 0.0193 | 266 | 10.2188 | 0.0762 |
|
320 |
+
| 10.2188 | 0.0194 | 267 | 10.2109 | 0.0762 |
|
321 |
+
| 10.2109 | 0.0195 | 268 | 10.2109 | 0.0763 |
|
322 |
+
| 10.2109 | 0.0195 | 269 | 10.2109 | 0.0762 |
|
323 |
+
| 10.2109 | 0.0196 | 270 | 10.2109 | 0.0761 |
|
324 |
+
| 10.2188 | 0.0197 | 271 | 10.2109 | 0.0761 |
|
325 |
+
| 10.2109 | 0.0198 | 272 | 10.2031 | 0.0760 |
|
326 |
+
| 10.2188 | 0.0198 | 273 | 10.2031 | 0.0761 |
|
327 |
+
| 10.2266 | 0.0199 | 274 | 10.2031 | 0.0762 |
|
328 |
+
| 10.2188 | 0.0200 | 275 | 10.2031 | 0.0762 |
|
329 |
+
| 10.2109 | 0.0201 | 276 | 10.1953 | 0.0761 |
|
330 |
+
| 10.2109 | 0.0201 | 277 | 10.1953 | 0.0762 |
|
331 |
+
| 10.1953 | 0.0202 | 278 | 10.1953 | 0.0762 |
|
332 |
+
| 10.2031 | 0.0203 | 279 | 10.1953 | 0.0763 |
|
333 |
+
| 10.2188 | 0.0203 | 280 | 10.1953 | 0.0765 |
|
334 |
+
| 10.1953 | 0.0204 | 281 | 10.1875 | 0.0766 |
|
335 |
+
| 10.1953 | 0.0205 | 282 | 10.1875 | 0.0767 |
|
336 |
+
| 10.2031 | 0.0206 | 283 | 10.1875 | 0.0767 |
|
337 |
+
| 10.1797 | 0.0206 | 284 | 10.1875 | 0.0766 |
|
338 |
+
| 10.1953 | 0.0207 | 285 | 10.1875 | 0.0765 |
|
339 |
+
| 10.1953 | 0.0208 | 286 | 10.1797 | 0.0764 |
|
340 |
+
| 10.1875 | 0.0209 | 287 | 10.1797 | 0.0764 |
|
341 |
+
| 10.1953 | 0.0209 | 288 | 10.1797 | 0.0765 |
|
342 |
+
| 10.1875 | 0.0210 | 289 | 10.1797 | 0.0765 |
|
343 |
+
| 10.1875 | 0.0211 | 290 | 10.1797 | 0.0768 |
|
344 |
+
| 10.1797 | 0.0211 | 291 | 10.1719 | 0.0770 |
|
345 |
+
| 10.1719 | 0.0212 | 292 | 10.1719 | 0.0771 |
|
346 |
+
| 10.1719 | 0.0213 | 293 | 10.1719 | 0.0772 |
|
347 |
+
| 10.1797 | 0.0214 | 294 | 10.1719 | 0.0773 |
|
348 |
+
| 10.1797 | 0.0214 | 295 | 10.1719 | 0.0773 |
|
349 |
+
| 10.1641 | 0.0215 | 296 | 10.1641 | 0.0773 |
|
350 |
+
| 10.1719 | 0.0216 | 297 | 10.1641 | 0.0773 |
|
351 |
+
| 10.1719 | 0.0217 | 298 | 10.1641 | 0.0773 |
|
352 |
+
| 10.1719 | 0.0217 | 299 | 10.1641 | 0.0773 |
|
353 |
+
| 10.1719 | 0.0218 | 300 | 10.1641 | 0.0773 |
|
354 |
+
| 10.1641 | 0.0219 | 301 | 10.1641 | 0.0773 |
|
355 |
+
| 10.1562 | 0.0219 | 302 | 10.1562 | 0.0772 |
|
356 |
+
| 10.1719 | 0.0220 | 303 | 10.1562 | 0.0771 |
|
357 |
+
| 10.1562 | 0.0221 | 304 | 10.1562 | 0.0772 |
|
358 |
+
| 10.1641 | 0.0222 | 305 | 10.1562 | 0.0773 |
|
359 |
+
| 10.1562 | 0.0222 | 306 | 10.1484 | 0.0773 |
|
360 |
+
| 10.1641 | 0.0223 | 307 | 10.1484 | 0.0773 |
|
361 |
+
| 10.1719 | 0.0224 | 308 | 10.1484 | 0.0775 |
|
362 |
+
| 10.1562 | 0.0224 | 309 | 10.1484 | 0.0775 |
|
363 |
+
| 10.1719 | 0.0225 | 310 | 10.1484 | 0.0775 |
|
364 |
+
| 10.1562 | 0.0226 | 311 | 10.1406 | 0.0774 |
|
365 |
+
| 10.1562 | 0.0227 | 312 | 10.1406 | 0.0774 |
|
366 |
+
| 10.1562 | 0.0227 | 313 | 10.1406 | 0.0773 |
|
367 |
+
| 10.1406 | 0.0228 | 314 | 10.1406 | 0.0774 |
|
368 |
+
| 10.1406 | 0.0229 | 315 | 10.1406 | 0.0774 |
|
369 |
+
| 10.1406 | 0.0230 | 316 | 10.1406 | 0.0774 |
|
370 |
+
| 10.1328 | 0.0230 | 317 | 10.1328 | 0.0775 |
|
371 |
+
| 10.1484 | 0.0231 | 318 | 10.1328 | 0.0775 |
|
372 |
+
| 10.1328 | 0.0232 | 319 | 10.1328 | 0.0775 |
|
373 |
+
| 10.1328 | 0.0232 | 320 | 10.1328 | 0.0775 |
|
374 |
+
| 10.125 | 0.0233 | 321 | 10.1328 | 0.0775 |
|
375 |
+
| 10.1406 | 0.0234 | 322 | 10.125 | 0.0776 |
|
376 |
+
| 10.1328 | 0.0235 | 323 | 10.125 | 0.0777 |
|
377 |
+
| 10.125 | 0.0235 | 324 | 10.125 | 0.0778 |
|
378 |
+
| 10.125 | 0.0236 | 325 | 10.125 | 0.0777 |
|
379 |
+
| 10.125 | 0.0237 | 326 | 10.125 | 0.0777 |
|
380 |
+
| 10.1328 | 0.0238 | 327 | 10.1172 | 0.0777 |
|
381 |
+
| 10.1172 | 0.0238 | 328 | 10.1172 | 0.0777 |
|
382 |
+
| 10.1172 | 0.0239 | 329 | 10.1172 | 0.0777 |
|
383 |
+
| 10.125 | 0.0240 | 330 | 10.1172 | 0.0778 |
|
384 |
+
| 10.1094 | 0.0240 | 331 | 10.1172 | 0.0778 |
|
385 |
+
| 10.1094 | 0.0241 | 332 | 10.1094 | 0.0777 |
|
386 |
+
| 10.1094 | 0.0242 | 333 | 10.1094 | 0.0776 |
|
387 |
+
| 10.1172 | 0.0243 | 334 | 10.1094 | 0.0775 |
|
388 |
+
| 10.125 | 0.0243 | 335 | 10.1094 | 0.0774 |
|
389 |
+
| 10.1172 | 0.0244 | 336 | 10.1094 | 0.0772 |
|
390 |
+
| 10.1016 | 0.0245 | 337 | 10.1016 | 0.0771 |
|
391 |
+
| 10.1094 | 0.0246 | 338 | 10.1016 | 0.0773 |
|
392 |
+
| 10.1172 | 0.0246 | 339 | 10.1016 | 0.0775 |
|
393 |
+
| 10.1094 | 0.0247 | 340 | 10.1016 | 0.0777 |
|
394 |
+
| 10.1172 | 0.0248 | 341 | 10.1016 | 0.0778 |
|
395 |
+
| 10.0938 | 0.0248 | 342 | 10.0938 | 0.0779 |
|
396 |
+
| 10.1016 | 0.0249 | 343 | 10.0938 | 0.0780 |
|
397 |
+
| 10.0938 | 0.0250 | 344 | 10.0938 | 0.0780 |
|
398 |
+
| 10.0938 | 0.0251 | 345 | 10.0938 | 0.0780 |
|
399 |
+
| 10.1016 | 0.0251 | 346 | 10.0938 | 0.0781 |
|
400 |
+
| 10.1094 | 0.0252 | 347 | 10.0859 | 0.0780 |
|
401 |
+
| 10.0938 | 0.0253 | 348 | 10.0859 | 0.0780 |
|
402 |
+
| 10.0938 | 0.0254 | 349 | 10.0859 | 0.0780 |
|
403 |
+
| 10.0859 | 0.0254 | 350 | 10.0859 | 0.0779 |
|
404 |
+
| 10.0859 | 0.0255 | 351 | 10.0859 | 0.0780 |
|
405 |
+
| 10.0938 | 0.0256 | 352 | 10.0781 | 0.0781 |
|
406 |
+
|
407 |
+
|
408 |
+
### Framework versions
|
409 |
+
|
410 |
+
- Transformers 4.41.2
|
411 |
+
- Pytorch 2.1.0a0+32f93b1
|
412 |
+
- Datasets 2.20.0
|
413 |
+
- Tokenizers 0.19.1
|
config.json
ADDED
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "openai-community/gpt2",
|
3 |
+
"activation_function": "gelu_new",
|
4 |
+
"architectures": [
|
5 |
+
"GPT2LMHeadModel"
|
6 |
+
],
|
7 |
+
"attn_pdrop": 0.1,
|
8 |
+
"bos_token_id": 50256,
|
9 |
+
"embd_pdrop": 0.1,
|
10 |
+
"eos_token_id": 50256,
|
11 |
+
"initializer_range": 0.02,
|
12 |
+
"layer_norm_epsilon": 1e-05,
|
13 |
+
"model_type": "gpt2",
|
14 |
+
"n_ctx": 1024,
|
15 |
+
"n_embd": 128,
|
16 |
+
"n_head": 2,
|
17 |
+
"n_inner": null,
|
18 |
+
"n_layer": 12,
|
19 |
+
"n_positions": 1024,
|
20 |
+
"reorder_and_upcast_attn": false,
|
21 |
+
"resid_pdrop": 0.1,
|
22 |
+
"scale_attn_by_inverse_layer_idx": false,
|
23 |
+
"scale_attn_weights": true,
|
24 |
+
"summary_activation": null,
|
25 |
+
"summary_first_dropout": 0.1,
|
26 |
+
"summary_proj_to_labels": true,
|
27 |
+
"summary_type": "cls_index",
|
28 |
+
"summary_use_proj": true,
|
29 |
+
"task_specific_params": {
|
30 |
+
"text-generation": {
|
31 |
+
"do_sample": true,
|
32 |
+
"max_length": 50
|
33 |
+
}
|
34 |
+
},
|
35 |
+
"torch_dtype": "float16",
|
36 |
+
"transformers_version": "4.41.2",
|
37 |
+
"use_cache": true,
|
38 |
+
"vocab_size": 50257
|
39 |
+
}
|
generation_config.json
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"bos_token_id": 50256,
|
4 |
+
"eos_token_id": 50256,
|
5 |
+
"transformers_version": "4.41.2"
|
6 |
+
}
|
logs/events.out.tfevents.1719842149.s_005_m.77188.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b724582005972baa62aedbca9d2cbd96118dc469d83142544f506b140c29a7cc
|
3 |
+
size 192074
|
merges.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ec7ea8bf59717d391ca0aa5d86b15de2c174bd19d301f695953cd23b95383805
|
3 |
+
size 30767592
|
special_tokens_map.json
ADDED
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": "<|endoftext|>",
|
3 |
+
"eos_token": "<|endoftext|>",
|
4 |
+
"unk_token": "<|endoftext|>"
|
5 |
+
}
|
tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer_config.json
ADDED
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_prefix_space": false,
|
3 |
+
"added_tokens_decoder": {
|
4 |
+
"50256": {
|
5 |
+
"content": "<|endoftext|>",
|
6 |
+
"lstrip": false,
|
7 |
+
"normalized": true,
|
8 |
+
"rstrip": false,
|
9 |
+
"single_word": false,
|
10 |
+
"special": true
|
11 |
+
}
|
12 |
+
},
|
13 |
+
"bos_token": "<|endoftext|>",
|
14 |
+
"clean_up_tokenization_spaces": true,
|
15 |
+
"eos_token": "<|endoftext|>",
|
16 |
+
"model_max_length": 1024,
|
17 |
+
"tokenizer_class": "GPT2Tokenizer",
|
18 |
+
"unk_token": "<|endoftext|>"
|
19 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b44364ed8f2463fe1cabeca16997785f0e1910269d517c23cc3ba6997bf93bea
|
3 |
+
size 6328
|
vocab.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|